How does cost growth affect decision making?

Additionally, the presented results in this article suggest that aircraft programs that have no cost growth from Contract Award to First Flight have a very good chance of reaching the end of DT&E without cost growth. In contrast, those programs that experience more than 10% cost growth by the end of DT&E have a very strong chance of remaining above 10% cost growth for the entirety of the program. This tangentially supports what Smith, White, Ritschel, and Thal (2016) mentioned in that having a solid and funded test plan often mitigates future cost growth. Conversely, lacking or possessing limited funds devoted to testing often has detrimental effects on the entire cost of the program as issues are discovered too late to fiscally rectify.

That is from Unmasking Cost Growth Behavior in the ARJ.

I’m not sure how these results support the idea that “a solid and funded test plan often mitigates future cost growth.” It sounds like if you were 10% over budget to first flight — and that reliably means you’ll be that much or over for the program — then engineering and requirements problems were the culprit and not testing.

It’s always hard for me to discern anything from cost growth and similar figures. Does it help us make incremental decisions? If my program was 10% over budget at first flight, how does that affect my plan? Should I slow-roll the program until engineering difficulties are resolved? Throw more resources to make schedule? Trade-off requirements? Cut the program all-together?

Even if cost growth told us something that a good manager on the ground didn’t already know, it often arrives late. Everyone’s always looking for early warning indicators.

I think programs become schedule driven, and skimp on testing, rather than event driven because milestone reviews and funding plans need to be aligned long before-the-fact. Changes to the timing of events, such as CDR, First Flight, OpEval, and so on, mean more briefings and explanations to the powers that be, because funding schedules are time-based even if the acquisition process can make itself event based.